HAZE-IMAGE-DATASET: A Large-Scale Benchmark for Image Dehazing in Variable Fog and Low-Light Conditions

To enhance image dehazing and visual recognition in real-world conditions, we introduce HAZE-IMAGE-DATASET, a large-scale dataset comprising nearly 42,000 images. It is constructed from 1,532 clean images sourced globally and captured using a Samsung smartphone, covering diverse natural and urban scenes. The dataset includes synthetic and real haze variations. Synthetic haze was generated using MATLAB-based atmospheric scattering models with depth maps for 10 fog levels. Colored haze was created using alpha blending (α = 0.4) in six colors: red, green, blue, yellow, white, and black. Low-light conditions were simulated via uniform darkening at 10 levels. Also, 616 real haze images were captured using a steam device to replicate genuine haze characteristics. Benchmarking was performed using MobileNetV2 and GoogLeNet. On a hazy subset, MobileNetV2 improved PSNR from 13.34 dB to 21.77 dB and SSIM from 0.805 to 0.950, Also, GoogleNet achieved a PSNR of 22.38 dB and SSIM of 0.953.Classification experiments using a seven-class subset showed high accuracy: 98.93% with AlexNet and 98.13% with GoogLeNet.The dataset is organized into haze types, levels, and color folders. It serves as a strong foundation for training and evaluating image dehazing and recognition models under different visibility conditions. It is available on GitHub: https://github.com/mustafa1jamal/HAZE-IMAGE-DATASET.